[Frontend] Do prompt_logprobs clamping for chat as well as completions (vllm-project#14225)

Signed-off-by: Harry Mellor <[email protected]>

Author: hmellor

vllm/entrypoints/openai/serving_chat.py

@@ -24,7 +24,8 @@
     RequestResponseMetadata, ToolCall, UsageInfo)
 from vllm.entrypoints.openai.reasoning_parsers import (ReasoningParser,
                                                        ReasoningParserManager)
-from vllm.entrypoints.openai.serving_engine import OpenAIServing
+from vllm.entrypoints.openai.serving_engine import (OpenAIServing,
+                                                    clamp_prompt_logprobs)
 from vllm.entrypoints.openai.serving_models import OpenAIServingModels
 from vllm.entrypoints.openai.tool_parsers import ToolParser, ToolParserManager
 from vllm.entrypoints.openai.tool_parsers.mistral_tool_parser import (
@@ -844,7 +845,7 @@ async def chat_completion_full_generator(
             model=model_name,
             choices=choices,
             usage=usage,
-            prompt_logprobs=final_res.prompt_logprobs,
+            prompt_logprobs=clamp_prompt_logprobs(final_res.prompt_logprobs),
         )
 
         return response

vllm/entrypoints/openai/serving_completion.py

@@ -23,7 +23,8 @@
                                               RequestResponseMetadata,
                                               UsageInfo)
 # yapf: enable
-from vllm.entrypoints.openai.serving_engine import OpenAIServing
+from vllm.entrypoints.openai.serving_engine import (OpenAIServing,
+                                                    clamp_prompt_logprobs)
 from vllm.entrypoints.openai.serving_models import OpenAIServingModels
 from vllm.logger import init_logger
 from vllm.outputs import RequestOutput
@@ -394,13 +395,7 @@ def request_output_to_completion_response(
         for final_res in final_res_batch:
             prompt_token_ids = final_res.prompt_token_ids
             assert prompt_token_ids is not None
-            prompt_logprobs = final_res.prompt_logprobs
-            if prompt_logprobs:
-                for logprob_dict in prompt_logprobs:
-                    if logprob_dict:
-                        for logprob_values in logprob_dict.values():
-                            if logprob_values.logprob == float('-inf'):
-                                logprob_values.logprob = -9999.0
+            prompt_logprobs = clamp_prompt_logprobs(final_res.prompt_logprobs)
             prompt_text = final_res.prompt
 
             token_ids: GenericSequence[int]

vllm/entrypoints/openai/serving_engine.py

@@ -42,7 +42,7 @@
 from vllm.pooling_params import PoolingParams
 from vllm.prompt_adapter.request import PromptAdapterRequest
 from vllm.sampling_params import BeamSearchParams, SamplingParams
-from vllm.sequence import Logprob
+from vllm.sequence import Logprob, PromptLogprobs
 from vllm.tracing import (contains_trace_headers, extract_trace_headers,
                           log_tracing_disabled_warning)
 from vllm.transformers_utils.tokenizer import AnyTokenizer, MistralTokenizer
@@ -535,3 +535,18 @@ def _get_model_name(self,
         if model_name is None:
             return self.models.base_model_paths[0].name
         return model_name
+
+
+def clamp_prompt_logprobs(
+    prompt_logprobs: Union[PromptLogprobs,
+                           None]) -> Union[PromptLogprobs, None]:
+    if prompt_logprobs is None:
+        return prompt_logprobs
+
+    for logprob_dict in prompt_logprobs:
+        if logprob_dict is None:
+            continue
+        for logprob_values in logprob_dict.values():
+            if logprob_values.logprob == float('-inf'):
+                logprob_values.logprob = -9999.0
+    return prompt_logprobs